GTX 1080ti periodic black screen crash with 100% fan speed.

Hello,

I'm having periodically occurring issues with my GPU, that I am unable to reproduce consistently.


The issue can be explained as this:
I play a game or do a benchmark - screens go black and the GPU fan ramps up to 100%
This seems like a complete disconnects off the GPU as the PC never recovers from the crash and the CPU is still responding to CAPS/NUM Lock and audio is still playing.

After rebooting when I start any intensive workload after the crash the computer will black screen again 98% of the time, like starting valley benchmark the computer will instantly backscreen when the screen shows the benchmark frames.

A lot of the time I can benchmark the GPU for hours and it will not crash until it suddenly crash again out of the blue, temperatures are completely fine under 60c at all times for all hardware.

The issue sometimes occurs during gaming sessions an MMO game that is majorly CPU intensive, GPU usage never exceeds 50% utilization, it also occurs during benchmarks and after uninstalling Windows and reinstalling the issue can be gone for 1-2months before it reappears out of the blue.

I've tried reinstalling Windows 10 with different builds, different drivers for all system hardware old and new drivers tested.
Completely dissembled the machine and checked all connections re-connecting everything and changed cables and had a visual inspection and a nose inspection to check for burnt electronics.

If disconnect the 1080ti and test with a 7950 and a 290 in crossfire or single handed I never get a single crash, I've ran both these cards stable for over 3days with 24/7 benchmarks.

I have 2 external monitors, 1x 2560x1440p 144Hz, DisplayPort.
1 2560x1080p 60Hz monitor, DVI


Event View occurrences:
Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.

The lastes event view error I got was: Unable to recover from a kernel exception. The application must close.
Error code: 3 (subcode 2)
(pid=8952 tid=1628 glz.exe 64bit)

Checking the dumpfile under: C:\Users\Brew\AppData\Local\CrashDumps

FurMark_ROG_Edition_x64.exe.8640:
Microsoft (R) Windows Debugger Version 10.0.18362.1 AMD64 Copyright (c) Micro - Pastebin.com

GLZ.exe.8952
Microsoft (R) Windows Debugger Version 10.0.18362.1 AMD64 Copyright (c) Micro - Pastebin.com


I don't know how to read and understand the dump files.

Let me know if there any more information necessary, I would really appreciate help in figuring out if the card is just bad, I have to RMA it but they will need to reproduce the issue and I need to figure out how they can do that.

Hi-

You posted in the Win 7 forum but you have Win 10. You can post this in the Tenforums but it might be more useful to post it in the NVIDIA forum. For instance, I found this in which the application is different but the exception code and .dll problem are the same-

https://devtalk.nvidia.com/default/t...-nvoglv64-dll/

However, only NVIDIA could evaluate if this might be applicable to you.

Hi,Thanks for the reply, I have also tried posting on the NVIDIA forum and contacting ASUS and NVIDIA directly, I thought the forum was also for Windows 10 my bad.

The issue seems to have been resolved after I connected the PCI-e 6+2 pin power cable to a 8pin adapter directly into the GPU.
This bypasses the +2pin cable entirely and seems like this has been the issue on separate cables. It is connected like this now: https://www.dhresource.com/0x0s/f2-a...dual-pci-e.jpg

No new event viewer issue connected to the GPU and I have not had a crash for 2 days, seems to have been a power delivery issue.

Excellent!

I'm glad you got it resolved.